MIT Develops a Robotic Assistant That Can Dress You
The researchers created a new algorithm that helps a robot find efficient motion plans to safely put clothing on its human counterpart.
Robots are great at performing specific tasks, such as lifting heavy objects and the precision assembly of small electronic components. Some jobs, though, are too difficult for robots as they require a level of spatial awareness and the ability to adapt to changing circumstances in their operational environment. Those problems become more complicated when robots interact with humans and work together to complete tasks. Engineers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) set out to conquer those issues with new insight into robotic movement, giving a robotic arm the ability to dress humans without harming them.
One of the challenges in creating a robot that can dress humans is its short field of vision, meaning it can't "see" a person's entire arm and, therefore, needs to calculate the position of the limb by focusing on a specific area – in this case, the elbow. This was accomplished using a state of estimation algorithm, which allows the robot to make a relatively precise educated guess on where the elbow is and how the arm is oriented. The algorithm takes the robot's measurement of the force applied to the cloth as input, then estimates the elbow's position by placing it in a box that encompasses all possible positions. The data then tells the robot how to move without harming the human.
"Our algorithm solves the general state estimation problem and could lend itself to many possible applications," explains Shen Li, a Ph.D. candidate in the MIT Department of Aeronautics and Astronautics. "The key to it all is having the ability to guess."
The team says the robotic arm could help the disabled perform tasks such as setting the kitchen table, assisting with cooking and other chores, and helping those who need mobility aid.